同心円錯視の測定に関する一考察

Mutually exclusive tendencies of over- and underestimation of the same circle under various conditions of measurement were reported in several studies of geometric optical illusion when concentric circles were used as stimuli. The purpose of the present study was to analyse experimentally those conditions that caused the contradictory results. Four experimental series were as follows: Exp. Series I was to see if over- and underestimation were really made at the same circle. The patterns of Fig. 6 were used as comparison stimuli in this series of experiment. Exp. Series II was designed to investigate the effects which the systematic change of arcs in comparison stimuli produced. The patterns of Fig. 7 were used as comparison stimuli. Exp. Series III and IV were to comparethe effects of horizontal vs. vertical constellation of comparison stimuli. The patterns of Fig. 11 were used. Table 2 gives the summarized resultsof Exp. Series I. A tendency of overestimation is observed when measrement is done by means of paired dots. The results of Exp. Series II are plotted in Fig. 8. A continuous change from overestimation to underestimation corresponding to a series of comparison stimuli is found. In Figs. 12 and 14, the results of Exp. Series III and IV are plotted respectively. Courses of change in these series are similar to that of Exp. Series II. The results of Alexander\u27s trend analysis are given in Tables 3, 4 and 5. Conclusions from above findings are: 1. The direction of illusion changes continuously from underestimation to overestimation as a complete circle is reduced to a fraction of it-arcs of various radian. 2. The rate of change is smaller where the tendency of underestimation is reported, while the gradient becomes steeper where there is overestimation. 3. Prominent individual differences are noted in the range where no illusion is noted